Concepts
The vocabulary, one page per idea.
Every concept page in this section follows the same five-section spine:
- What it is. One paragraph definition.
- Invariants. The things that are always true about it.
- What it refuses. Canonical error strings.
- What it doesn’t do. Things that surprise newcomers.
- See also. Cross-links.
| Concept | One-line summary |
|---|---|
| Typed images | The container family the API hands you — NeuroVol, NeuroVec, NeuroHyperVec. |
| NeuroSpace | The frozen spatial frame that travels with every typed image. |
| Receipt | Content-addressable provenance — what’s hashed, what isn’t, why. |
| Provenance | How .provenance is produced, chained, and rehydrated across the file boundary. |
| VoxelSeriesStore | The one protocol behind dense, sparse, mapped, and file-backed 4-D containers. |
| ROI result | The ROIExtractionResult / SearchlightResult family. |
| Contract failures | The closed set of canonical refusals the API raises. |
| Atlas | Typed parcellation as an analysis input. |
The first time you read these, read top-to-bottom. After that, treat the pages as glossary entries — every other doc page links here instead of redefining vocabulary.
Need the inverse — concept → API symbols → recipe → spec, in one table? See the Concept ↔︎ API map.
Back to top